Semiconductor News 20260130

Main News

1. TSMC Restructures Advanced Packaging Strategy to Maintain Production Capacity in Taiwan
   TSMC is restructuring its advanced packaging (CoWoS) expansion strategy to meet surging demand for AI and is reorganizing plans for AP8, AP7, and its US plants. NVIDIA CEO Jensen Huang also denied rumors that 40% of Taiwan’s manufacturing capacity would be moved to the US, stating that it would be for new expansion. This indicates that the company intends to optimize its global production structure while maintaining its manufacturing base in Taiwan.
2. Samsung Electronics Fights Back Against NVIDIA with HBM4, Q4 Profit Surges
   Samsung Electronics has reclaimed the top spot in memory sales, with operating profit for the fourth quarter of 2025 topping 20 trillion won on the back of recovering memory demand. In addition, the company’s next-generation HBM4 memory has reportedly been certified and adopted by NVIDIA, launching a counterattack in the HBM market, where SK Hynix is ahead. Meanwhile, the concentration of production on AI chips has increased supply pressure for smartphones and displays.
3. Micron Begins Construction of New $24 Billion NAND Plant in Singapore
   Micron has begun construction of a new advanced wafer fabrication facility in Singapore to meet AI and data center demand. The project will cost $24 billion over the next 10 years. In addition, NVIDIA CEO Jensen Huang has endorsed Micron as the AI Memory Alliance in Taiwan, simultaneously strengthening production capacity and partnerships in Asia.
4. Alibaba Announces New AI Chip to Compete with NVIDIA and Observes IPO
   T-Head, the semiconductor division of Alibaba Group, has announced a new AI chip that is said to rival the performance of NVIDIA’s H20 and A100. By doing so, the company aims to capture alternative demand in the Chinese market, which is under U.S. export restrictions. In addition, the market is increasingly speculating on a spin-off and IPO (initial public offering) of the division, a move that would accelerate the development and competitiveness of AI semiconductors through independent financing.
5. ASML Posts Strong Results but Plans to Cut 1,700 Jobs to Improve Efficiency
   Semiconductor equipment giant ASML has posted strong fourth-quarter and full-year 2025 results, but has announced plans to cut 1,700 jobs to improve organizational efficiency. demand for AI has spread to the manufacturing floor, and the company is planning to reduce its workforce by 1,700 workers. Orders for EUV lithography equipment and other products are strong, but the company is expected to review its cost structure to prepare for future market volatility and decisively make strategic adjustments to create a foundation for sustainable growth.
6. ‘Gray Rhino’ Risk of Rising Semiconductor Prices Emerges in Chinese Market
   A rising price spiral has begun in China’s semiconductor market, threatening to become a “gray Rhino” (a probable and significant but often overlooked risk) for the supply chain in 2026. due to the AI boom. Inflationary pressures are increasing, triggered by overheated demand and shortages of certain chips. There are concerns that this trend will spill over into procurement costs and product prices in a wide range of industries, including smartphones and automobiles.
7. Toyota, Suzuki to Adopt Chinese SoCs in Shift in Automotive Semiconductor Strategy
   Japanese automakers such as Toyota Motor Corporation and Suzuki Motor Corporation are moving to adopt Chinese-made system-on-chips (SoCs) as a turning point in their automotive semiconductor strategy. By adopting the technologies of Chinese suppliers, which are ahead of the competition in electric vehicles (EV) and automated driving technologies and are cost competitive, they are said to be aiming to maintain and strengthen their competitiveness in the global market, and the shift from the conventional supply chain is attracting attention.
8. SK H ynix Maintains Dominance in AI Memory Market, but Warns of Supply Shortages in 2026
   SK Hynix is extending its profit lead in the HBM market for AI memory, but expects DRAM supply shortages to continue in the second half of 2026 and similar tightness in NAND flash The company is also forecasting similar shortages for NAND flash in the second half of 2026. The company is strengthening its strategy to lock in customers to its AI memory products, and its focus on next-generation products such as HBM4 could lead to supply constraints in general-purpose memory.
9. Photoelectric fusion technology is accelerating, NVIDIA estimates that it can reduce power consumption by 70%.
   Photoelectric fusion (CPO) technology, which converts electrical signals into optical signals, is attracting increasing attention as a trump card for reducing power consumption in data centers. With the development of this technology by Japanese companies such as QD Laser and the joint development of Lightmatter and GUC, it is becoming a key technology to support the sustainability of AI infrastructures worldwide.
10. UMC Strengthens Advanced Packaging and Silicon Photonics to Meet AI Demand
    Taiwanese foundry UMC is accelerating its development of advanced packaging technology and silicon photonics to meet growing AI demand. The company expects momentum in mature nodes such as the 22nm process to continue through 2026, and its strategy is to capture demand for AI edge devices and telecom infrastructure through enhancements in specialty processes. not only TSMC, but also mid-sized foundries are quickly adapting to the AI market.

News List